Semaphore详解

AtomicInteger解析：http://donald-draper.iteye.com/blog/2359555
锁持有者管理器AbstractOwnableSynchronizer：http://donald-draper.iteye.com/blog/2360109
AQS线程挂起辅助类LockSupport：http://donald-draper.iteye.com/blog/2360206
AQS详解-CLH队列，线程等待状态：http://donald-draper.iteye.com/blog/2360256
AQS-Condition详解：http://donald-draper.iteye.com/blog/2360381
可重入锁ReentrantLock详解：http://donald-draper.iteye.com/blog/2360411
CountDownLatch使用场景：http://donald-draper.iteye.com/blog/2348106
CountDownLatch详解：http://donald-draper.iteye.com/blog/2360597
CyclicBarrier详解：http://donald-draper.iteye.com/blog/2360812

/*
 * ORACLE PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
 * Written by Doug Lea with assistance from members of JCP JSR-166
 * Expert Group and released to the public domain, as explained at
 * http://creativecommons.org/publicdomain/zero/1.0/
 */

package java.util.concurrent;
import java.util.*;
import java.util.concurrent.locks.*;
import java.util.concurrent.atomic.*;

/**
 * A counting semaphore.  Conceptually, a semaphore maintains a set of
 * permits.  Each {@link #acquire} blocks if necessary until a permit is
 * available, and then takes it.  Each {@link #release} adds a permit,
 * potentially releasing a blocking acquirer.
 * However, no actual permit objects are used; the {@code Semaphore} just
 * keeps a count of the number available and acts accordingly.
 *
 一个计算的信号量，维持着一个许可集。如果许可集中，无许可，线程acquire，
 将会阻塞，直到其他线程释放许可。线程每一次释放#release，则添加一个许可，潜在地
 释放一个阻塞信号获取者。信号量的许可，实际上并不是一对象，仅仅保证一定数量的虚拟许可证。
 * <p>Semaphores are often used to restrict the number of threads than can
 * access some (physical or logical) resource. For example, here is
 * a class that uses a semaphore to control access to a pool of items:
 * <pre>
 信号量经常被用于，只有一定数量的线程访问一些物理或逻辑资源。比如用信号量控制池对象的获取。
 * class Pool {
 *   private static final int MAX_AVAILABLE = 100;
 *   private final Semaphore available = new Semaphore(MAX_AVAILABLE, true);
 *
 *   public Object getItem() throws InterruptedException {
 *     available.acquire();
 *     return getNextAvailableItem();
 *   }
 *
 *   public void putItem(Object x) {
 *     if (markAsUnused(x))
 *       available.release();
 *   }
 *
 *   // Not a particularly efficient data structure; just for demo
 *
 *   protected Object[] items = ... whatever kinds of items being managed
 *   protected boolean[] used = new boolean[MAX_AVAILABLE];
 *
 *   protected synchronized Object getNextAvailableItem() {
 *     for (int i = 0; i < MAX_AVAILABLE; ++i) {
 *       if (!used[i]) {
 *          used[i] = true;
 *          return items[i];
 *       }
 *     }
 *     return null; // not reached
 *   }
 *
 *   protected synchronized boolean markAsUnused(Object item) {
 *     for (int i = 0; i < MAX_AVAILABLE; ++i) {
 *       if (item == items[i]) {
 *          if (used[i]) {
 *            used[i] = false;
 *            return true;
 *          } else
 *            return false;
 *       }
 *     }
 *     return false;
 *   }
 *
 * }
 * </pre>
 *
 * <p>Before obtaining an item each thread must acquire a permit from
 * the semaphore, guaranteeing that an item is available for use. When
 * the thread has finished with the item it is returned back to the
 * pool and a permit is returned to the semaphore, allowing another
 * thread to acquire that item.  Note that no synchronization lock is
 * held when {@link #acquire} is called as that would prevent an item
 * from being returned to the pool.  The semaphore encapsulates the
 * synchronization needed to restrict access to the pool, separately
 * from any synchronization needed to maintain the consistency of the
 * pool itself.
 *
在线程从对象池，获取对象前，必须从信号量获取许可，用于保证对象时可利用的。
当线程任务完成时，对象将会被放回池中，释放许可，同时允许其他线程获取对象。
如果线程acquire，但没有持有同步锁，则对象将返回池中。信号量中的同步器需要严格的
控制对象池的访问，与其他维持对象池一致性的同步器相互独立。


 * <p>A semaphore initialized to one, and which is used such that it
 * only has at most one permit available, can serve as a mutual
 * exclusion lock.  This is more commonly known as a [i]binary
 * semaphore[/i], because it only has two states: one permit
 * available, or zero permits available.  When used in this way, the
 * binary semaphore has the property (unlike many {@link Lock}
 * implementations), that the "lock" can be released by a
 * thread other than the owner (as semaphores have no notion of
 * ownership).  This can be useful in some specialized contexts, such
 * as deadlock recovery.
 *
当信号量被初始化为1时，作为互斥锁，可以用于最多只有一个 permit可以用的场景。
这种方式比较有名的一种是二进制信号量，因为它只有两种状态，1表示可利用，0表示
无permits可利用。二进制信号量，有一个属性，锁可以被其他非持有锁的线程释放。
这种特性在一些特殊的上下文场景中，比较拥有，比如恢复死锁。


 * <p> The constructor for this class optionally accepts a
 * [i]fairness[/i] parameter. When set false, this class makes no
 * guarantees about the order in which threads acquire permits. In
 * particular, [i]barging[/i] is permitted, that is, a thread
 * invoking {@link #acquire} can be allocated a permit ahead of a
 * thread that has been waiting - logically the new thread places itself at
 * the head of the queue of waiting threads. When fairness is set true, the
 * semaphore guarantees that threads invoking any of the {@link
 * #acquire() acquire} methods are selected to obtain permits in the order in
 * which their invocation of those methods was processed
 * (first-in-first-out; FIFO). Note that FIFO ordering necessarily
 * applies to specific internal points of execution within these
 * methods.  So, it is possible for one thread to invoke
 * {@code acquire} before another, but reach the ordering point after
 * the other, and similarly upon return from the method.
 * Also note that the untimed {@link #tryAcquire() tryAcquire} methods do not
 * honor the fairness setting, but will take any permits that are
 * available.
 *
 信号量的构造函数，中带一个公平性参数。当设置为false时，信号量不能够保证，线程
 能够按顺序获取许可。在特殊情况下，barging是允许的，由于一个新线程将自己放在队列的
 头部，当调用acquire时，可能会在已经等待线程，前面获取许可。当为公平true时，
 信号量保证，线程按照，他们获取信号的顺序，给予许可。FIFO队列中，并不能保证却对的
 顺序，一个线程可能调用获取信号量，在另一线程前面，但另一个线程先到达顺序点。
 tryAcquire方法，不能保证绝对的公平性， 当许可可利用，则许可将被分配。



 * <p>Generally, semaphores used to control resource access should be
 * initialized as fair, to ensure that no thread is starved out from
 * accessing a resource. When using semaphores for other kinds of
 * synchronization control, the throughput advantages of non-fair
 * ordering often outweigh fairness considerations.
 *
 信号量，用于控制资源的访问时，应该初始化为公平锁，以保证不会有线程，几乎
 访问不到资源。当信号量用于其他场景时，非公平锁，可以提高吞吐量。

 * <p>This class also provides convenience methods to {@link
 * #acquire(int) acquire} and {@link #release(int) release} multiple
 * permits at a time.  Beware of the increased risk of indefinite
 * postponement when these methods are used without fairness set true.


 信号量允许一次获取或释放多个信号量。当这些方法以非公平锁的方式使用，将会
 增加不确定性的风险
 *
 * <p>Memory consistency effects: Actions in a thread prior to calling
 * a "release" method such as {@code release()}
 * [url=package-summary.html#MemoryVisibility]<i>happen-before</i>[/url]
 * actions following a successful "acquire" method such as {@code acquire()}
 * in another thread.
 *
 内存一致性：一个线程释放锁动作，发生在另一个线程成功获取锁的前面。
 * @since 1.5
 * @author Doug Lea
 *
 */

public class Semaphore implements java.io.Serializable {
    private static final long serialVersionUID = -3222578661600680210L;
    /** All mechanics via AbstractQueuedSynchronizer subclass */
    //内部同步锁，基于AQS实现
    private final Sync sync;

    /**
     * Synchronization implementation for semaphore.  Uses AQS state
     * to represent permits. Subclassed into fair and nonfair
     * versions.
     */
    abstract static class Sync extends AbstractQueuedSynchronizer {
        private static final long serialVersionUID = 1192457210091910933L;
        //以锁的状态来，来存储许可
        Sync(int permits) {
            setState(permits);
        }

        final int getPermits() {
            return getState();
        }
        //非公平方式获取锁
        final int nonfairTryAcquireShared(int acquires) {
            for (;;) {
                int available = getState();
                int remaining = available - acquires;
		//如果，无许可可用，则返回，有，则CAS更新锁状态
                if (remaining < 0 ||
                    compareAndSetState(available, remaining))
                    return remaining;
            }
        }
        //释放共享锁
        protected final boolean tryReleaseShared(int releases) {
            for (;;) {
	        //释放的信号量，不能大于当前可用许可
                int current = getState();
                int next = current + releases;
                if (next < current) // overflow
                    throw new Error("Maximum permit count exceeded");
		 //CAS更新锁状态
                if (compareAndSetState(current, next))
                    return true;
            }
        }
        //减少当前可用的许可
        final void reducePermits(int reductions) {
            for (;;) {
                int current = getState();
                int next = current - reductions;
                if (next > current) // underflow
                    throw new Error("Permit count underflow");
                if (compareAndSetState(current, next))
                    return;
            }
        }
        /**
     * Acquires and returns all permits that are immediately available.
     *获取返回当前以及可用的许可
     * @return the number of permits acquired
     */
        final int drainPermits() {
            for (;;) {
                int current = getState();
                if (current == 0 || compareAndSetState(current, 0))
                    return current;
            }
        }
    }
     /**
     * NonFair version，非公平锁
     */
    static final class NonfairSync extends Sync {
        private static final long serialVersionUID = -2694183684443567898L;

        NonfairSync(int permits) {
            super(permits);
        }

        protected int tryAcquireShared(int acquires) {
            return nonfairTryAcquireShared(acquires);
        }
    }

    /**
     * Fair version，公平锁
     */
    static final class FairSync extends Sync {
        private static final long serialVersionUID = 2014338818796000944L;

        FairSync(int permits) {
            super(permits);
        }

        protected int tryAcquireShared(int acquires) {
            for (;;) {
	        //先看有没有前驱，有则返回，获取信号失败，没有前驱，则尝试获取信号锁
                if (hasQueuedPredecessors())
                    return -1;
                int available = getState();
                int remaining = available - acquires;
                if (remaining < 0 ||
                    compareAndSetState(available, remaining))
                    return remaining;
            }
        }
    }
    //默认为非公平锁，许可必须为正值
    public Semaphore(int permits) {
        sync = new NonfairSync(permits);
    }
    //带公平性参数的信号量，构造
     public Semaphore(int permits, boolean fair) {
        sync = fair ? new FairSync(permits) : new NonfairSync(permits);
    }
}

尝试获取锁，可中断

public void acquire() throws InterruptedException {
        sync.acquireSharedInterruptibly(1);
    }

//AQS

public final void acquireSharedInterruptibly(int arg)
            throws InterruptedException {
        if (Thread.interrupted())
	    //如果中断，则抛出中断异常
            throw new InterruptedException();
	    //如果获取失败，则自旋
        if (tryAcquireShared(arg) < 0)
            doAcquireSharedInterruptibly(arg);
    }
//待子类扩展
 protected int tryAcquireShared(int arg) {
        throw new UnsupportedOperationException();
    }

tryAcquireShared放在AQS中为空体，实际为信号量中的内部SYNC中的方法，上面
我们已经看过。这个方法我们在前面有说过，这里简单说一下，当尝试获取锁，失败，添加到等待队列自旋等待，尝试获取锁。
//AQS

 private void doAcquireSharedInterruptibly(int arg)
        throws InterruptedException {
        final Node node = addWaiter(Node.SHARED);
        boolean failed = true;
        try {
            for (;;) {
                final Node p = node.predecessor();
                if (p == head) {
                    int r = tryAcquireShared(arg);
                    if (r >= 0) {
                        setHeadAndPropagate(node, r);
                        p.next = null; // help GC
                        failed = false;
                        return;
                    }
                }
                if (shouldParkAfterFailedAcquire(p, node) &&
                    parkAndCheckInterrupt())
                    throw new InterruptedException();
            }
        } finally {
            if (failed)
                cancelAcquire(node);
        }
    }

以不可中断方式，获取共享锁

public void acquireUninterruptibly() {
        sync.acquireShared(1);
    }

//AQS
 

 /**
     * Acquires in shared uninterruptible mode.
     * @param arg the acquire argument
     */
已共享非中断模式，获取锁
 public final void acquireShared(int arg) {
        if (tryAcquireShared(arg) < 0)
            doAcquireShared(arg);
    }
      private void doAcquireShared(int arg) {
        final Node node = addWaiter(Node.SHARED);
        boolean failed = true;
        try {
            boolean interrupted = false;
            for (;;) {
                final Node p = node.predecessor();
                if (p == head) {
                    int r = tryAcquireShared(arg);
                    if (r >= 0) {
                        setHeadAndPropagate(node, r);
                        p.next = null; // help GC
                        if (interrupted)
			    //关键在这，如果中断，则自中断，消除中断位
                            selfInterrupt();
                        failed = false;
                        return;
                    }
                }
                if (shouldParkAfterFailedAcquire(p, node) &&
                    parkAndCheckInterrupt())
                    interrupted = true;
            }
        } finally {
            if (failed)
                cancelAcquire(node);
        }
    }
  /**
     * Convenience method to interrupt current thread.
     */
    private static void selfInterrupt() {
        Thread.currentThread().interrupt();
    }

//Semaphore
尝试获取锁时，是以非公平的方式，抢占锁
 

  public boolean tryAcquire() {
        return sync.nonfairTryAcquireShared(1) >= 0;
    }

尝试获取共享锁，当超时，还没有获取锁，则取消锁的获取

 public boolean tryAcquire(long timeout, TimeUnit unit)
        throws InterruptedException {
        return sync.tryAcquireSharedNanos(1, unit.toNanos(timeout));
    }

以可中断方式，获取permits个许可

 public void acquire(int permits) throws InterruptedException {
        if (permits < 0) throw new IllegalArgumentException();
        sync.acquireSharedInterruptibly(permits);
    }

以非可中断方式，获取permits个许可

public void acquireUninterruptibly(int permits) {
        if (permits < 0) throw new IllegalArgumentException();
        sync.acquireShared(permits);
    }

以非公平方式，尝试获取permits个许可

public boolean tryAcquire(int permits) {
        if (permits < 0) throw new IllegalArgumentException();
        return sync.nonfairTryAcquireShared(permits) >= 0;
    }

尝试获取permits个许可，当超时，还没有获取锁，则取消锁的获取

  public boolean tryAcquire(int permits, long timeout, TimeUnit unit)
        throws InterruptedException {
        if (permits < 0) throw new IllegalArgumentException();
        return sync.tryAcquireSharedNanos(permits, unit.toNanos(timeout));
    }

//释放锁

 public void release() {
        sync.releaseShared(1);
    }

//AQS

 public final boolean releaseShared(int arg) {
        if (tryReleaseShared(arg)) {
            doReleaseShared();
            return true;
        }
        return false;
    }
//唤醒后继节点线程
  * Release action for shared mode -- signal successor and ensure
     * propagation. (Note: For exclusive mode, release just amounts
     * to calling unparkSuccessor of head if it needs signal.)
     */
    private void doReleaseShared() {
        /*
         * Ensure that a release propagates, even if there are other
         * in-progress acquires/releases.  This proceeds in the usual
         * way of trying to unparkSuccessor of head if it needs
         * signal. But if it does not, status is set to PROPAGATE to
         * ensure that upon release, propagation continues.
         * Additionally, we must loop in case a new node is added
         * while we are doing this. Also, unlike other uses of
         * unparkSuccessor, we need to know if CAS to reset status
         * fails, if so rechecking.
         */
        for (;;) {
            Node h = head;
            if (h != null && h != tail) {
                int ws = h.waitStatus;
                if (ws == Node.SIGNAL) {
                    if (!compareAndSetWaitStatus(h, Node.SIGNAL, 0))
                        continue;            // loop to recheck cases
                    unparkSuccessor(h);
                }
                else if (ws == 0 &&
                         !compareAndSetWaitStatus(h, 0, Node.PROPAGATE))
                    continue;                // loop on failed CAS
            }
            if (h == head)                   // loop if head changed
                break;
        }
    }

//Semaphore
//释放permits个许可

 public void release(int permits) {
        if (permits < 0) throw new IllegalArgumentException();
        sync.releaseShared(permits);
    }
  /**
     * Returns the current number of permits available in this semaphore.
     *
     * <p>This method is typically used for debugging and testing purposes.
     *
     * @return the number of permits available in this semaphore
     */
    //当前可用许可
    public int availablePermits() {
        return sync.getPermits();
    }

    /**
     * Acquires and returns all permits that are immediately available.
     *获取返回当前以及可用的许可
     * @return the number of permits acquired
     */
    public int drainPermits() {
        return sync.drainPermits();
    }

    /**
     * Shrinks the number of available permits by the indicated
     * reduction. This method can be useful in subclasses that use
     * semaphores to track resources that become unavailable. This
     * method differs from {@code acquire} in that it does not block
     * waiting for permits to become available.
     * 减少当前可用的许可，这个方法在子类想要用信号量，追踪资源是否可用是
     非常有用，此方法不会阻塞。
     * @param reduction the number of permits to remove
     * @throws IllegalArgumentException if {@code reduction} is negative
     */
    protected void reducePermits(int reduction) {
        if (reduction < 0) throw new IllegalArgumentException();
        sync.reducePermits(reduction);
    }

    /**
     * Returns {@code true} if this semaphore has fairness set true.
     *
     * @return {@code true} if this semaphore has fairness set true
     */
    public boolean isFair() {
        return sync instanceof FairSync;
    }

    /**
     * Queries whether any threads are waiting to acquire. Note that
     * because cancellations may occur at any time, a {@code true}
     * return does not guarantee that any other thread will ever
     * acquire.  This method is designed primarily for use in
     * monitoring of the system state.
     *
     * @return {@code true} if there may be other threads waiting to
     *         acquire the lock
     */
    public final boolean hasQueuedThreads() {
        return sync.hasQueuedThreads();
    }

    /**
     * Returns an estimate of the number of threads waiting to acquire.
     * The value is only an estimate because the number of threads may
     * change dynamically while this method traverses internal data
     * structures.  This method is designed for use in monitoring of the
     * system state, not for synchronization control.
     *
     * @return the estimated number of threads waiting for this lock
     */
    public final int getQueueLength() {
        return sync.getQueueLength();
    }

    /**
     * Returns a collection containing threads that may be waiting to acquire.
     * Because the actual set of threads may change dynamically while
     * constructing this result, the returned collection is only a best-effort
     * estimate.  The elements of the returned collection are in no particular
     * order.  This method is designed to facilitate construction of
     * subclasses that provide more extensive monitoring facilities.
     *
     * @return the collection of threads
     */
    protected Collection<Thread> getQueuedThreads() {
        return sync.getQueuedThreads();
    }

总结：

信号量，维持着一个许可集。如果许可集中，无许可，线程acquire，
将会阻塞，直到其他线程释放许可。线程每一次释放#release，则添加一个许可，潜在地
释放一个阻塞信号获取者。信号量的许可，实际上并不是一对象，仅仅保证一定数量的虚拟许可证。信号量经常被用于，只有一定数量的线程访问一些物理或逻辑资源。比如用信号量控制池对象的获取。当信号量被初始化为1时，作为互斥锁，可以用于最多只有一个 permit可以用的场景。这种方式比较有名的一种是二进制信号量，因为它只有两种状态，1表示可利用，0表示无permits可利用。二进制信号量，有一个属性，锁可以被其他非持有锁的线程释放。
这种特性在一些特殊的上下文场景中，比较拥有，比如恢复死锁。信号量中的锁有公平和非公平方式，这个和我们前面讲的可重入锁，有点相似。非公平方式，获取锁，首先检查锁是否可用，可用则获取，公平锁获取锁，先检查有没有前驱节点，有则等待，没有则获取。获取锁的方法，有多种，有公平的和非公平，则个要看需求。